Prepared by Erica H. Dunn for the Extensive Monitoring Technical Committee of the Migration Monitoring Council June, 1995Direct comments to: Erica H. Dunn Canadian Wildlife Service National Wildlife Research Centre 100 Gamelin Blvd. Hull, Quebec, Canada K1A 0H3 Tel: (819) 994-0182 FAX: (819) 953-6612 e-mail: email@example.com Additional copies available from the same source, or from Greg Butcher, American Birding Association, P.O. Box 6599, Colorado Springs, CO 80934-6599
PREFACE In September 1993, a workshop was held to evaluate the potential of counting birds during migration as a means of assessing population change in landbirds (organized by the Canadian Wildlife Service and the U.S. Fish and Wildlife Service, now the National Biological Service). The workshop proposed formation of a "Migration Monitoring Council" to implement its recommendations (Blancher et al. 1993). In March, 1994 the Council met and appointed two technical committees charged with establishing standards and guidelines for operations of monitoring programs. The "Intensive Committee" is responsible for requirements of intensively operated sites such as bird observatories and banding stations. The "Extensive Committee" was charged with developing guidelines for geographically- dispersed migration counts, specifically checklist programs. This document presents the Extensive Committee's recommendation.
MIGRATION MONITORING COUNCIL AND TECHNICAL COMMITTEES Migration Monitoring Council *Peter Blancher, CWS, National Wildlife Research Centre Michael Bradstreet, Long Point Bird Observatory Greg Butcher, American Birding Association André Cyr, University of Sherbrooke Loney Dickson, CWS, Edmonton *Sam Droege, U.S. National Biological Service Bill Murphy, Ottawa Banding Group Nadav Nur, Point Reyes Bird Observatory C.J. Ralph, U.S. Forest Service Stan Temple, University of Wisconsin Intensive Sites Technical Committee John Hagan III, Manomet Bird Observatory Keith Hobson, CWS, Saskatoon *David Hussell, Ontario Ministry of Natural Resources Nadav Nur, Point Reyes Bird Observatory *C.J. Ralph, U.S. Forest Service Extensive Technical Committee *Greg Butcher, American Birding Association Jim Cox, Florida Fish and Game Brenda Dale, CWS, Edmonton *Erica Dunn, CWS, National Wildlife Research Centre Jeff Price, U.S. Nat. Biological Service, North Dakota Ken Rosenberg, Cornell Laboratory of Ornithology Rick West, Delaware Spring Bird Count * = Co-chairs of Council or Committee CWS = Canadian Wildlife Service, Environment Canada
Checklists are pre-printed lists of species on which observers can record their observations for an area of any size, during an observation period of any length. Some people keep year lists; others fill in a list every time they go birding. Some record only that a species was seen (the "check" of the checklist) whereas others record numbers of each species detected. Throughout this document we use the term "checklist" for all such lists, whether or not bird numbers are recorded. Although we recommend recording numbers (see below), "checklist" is a widespread, generic term that is likely to remain in general use. Rather than define a new term, we hope to redefine the concept of what a checklist should be.
Although bird observers have compiled lists for centuries, systematic record-keeping only became widespread in the last several decades. Development of birding tools such as field guides and relatively inexpensive binoculars helped bird- watching expand from the near-profession of a few to the recreation of tens of thousands (Davis 1994). The advent of the computer age has provided new opportunities for constructive use of pooled data, because now we can quickly find and compile results by species, location, year or other parameters. As Internet connections grow, birders should be able to submit data electronically, making it more feasible than ever to collate data contributed by large numbers of observers from broad geographic areas.
A few regional checklist programs have pioneered in taking advantage of these developments. Quebec's 'ÉPOQ' program ('Étude des Populations d'Oiseaux du Quebec', or Population Studies of Quebec's Birds) began in the 1950's, and currently computerizes about 10,000 checklists annually (Cyr and Larivée 1993). The Wisconsin Checklist Program has compiled weekly checklists since 1982 (Rolley 1994). There are also some long- standing programs that involve large numbers of people in counting migrants on a single day in May (e.g. the Delaware spring bird count, West 1992).
As the capability for running cooperative checklist programs has become easier, their numbers have shown signs of growing (mainly through interest by state or provincial organizations). Now is the time to consider the common goals of all such projects and to recommend guidelines that will make results compatible and comparable across North America.
In developing the guidelines that follow, we began with several basic considerations. First, the systematic collection of checklist data will be most successful if the same lists can be used in every season and for any species in the organizing group's coverage area. Although the authors of this document are particularly interested in data collected during migration, it would seem confusing and counter-productive to have different checklist programs for different seasons. The guidelines we propose are general ones that apply to any time of year.
A related consideration is that wide participation is a key to success. Checklists should therefore be designed to collect information that birders are interested in contributing. Lists should not be tedious to complete, should not ask for irrelevant information and should not require behavior much different from normal birding practice. Checklist programs should, however, collect data that will allow a degree of post-hoc standardization.
Finally, the data collected should be straightforward enough for simple analysis by people who do not have great statistical expertise, while at the same time including the elements required for more sophisticated treatment. Data should be compiled in a format that is compatible across data- collection programs, to promote ease of pooled analysis.
Any program that requires as much time and effort as a cooperative checklist project does should have a good reason for existing. Clarity of goals also helps us in developing recommended guidelines, as the data collected must be of a nature and quality that can address the questions we wish to answer.
The most common uses of checklist data in the past have been to determine whether, and when, a species is present at a given locality. Examples range from local bird status reports to regional ones (e.g. David 1980, Temple and Cary 1987, Cyr and Larivée 1995). State and provincial atlases of breeding birds are based on a specialized form of checklist data. The timing of migration can be determined, as well as the speed at which migrants move between their summer and winter homes. For remote regions that have been little explored, checklist data collected by travellers may be a sole source of biological inventory; one that gives clues to other flora and fauna that are likely to be present along with the birds. In such regions it may be a better use of resources to establish a checklist program than to concentrate on more intensive types of monitoring that can only cover a few sites.
Checklist information in regional data bases can be analyzed to indicate which geographic areas or habitats are the most species rich (e.g. Taylor and Smith 1987), or to pinpoint localities that are home to rare or endangered species. Results of such analyses are frequently used in environmental assessment (Dance and Fraser 1987).
Checklists can also be used to detect changes in abundance. For example, analyses of ÉPOQ data (Cyr and Larivée 1993, Dunn and Larivée in prep.) have indicated that long-term trends in checklist data parallel the changes detected by the Breeding Bird Survey, an independent program with a standardized sampling procedure (Peterjohn 1994). Agreement is not total (Dunn and Hussell 1995), and the success of these comparisons may depend in part on the very large ÉPOQ samples available on a daily basis (Dunn and Larivée in prep.). If birders change their behavior over time (for example, abandoning sites as they lose birds in order to visit better ones), then important trends could go undocumented. However, careful analysis can overcome some short-comings of the data (such as limiting analysis to data from a standard set of sites, or to counts of some minimum duration and quality).
An established checklist program would be a logical repository for historic data, making it available for future use instead of being lost. (See, for example, Hill and Hagan 1991, in which population trends since 1937 were inferred from 2 birders' field records).
Standardized monitoring programs with statistically- justified sampling protocols will always give more precise population trend estimates than checklist data. Nonetheless, we should not be deterred from taking advantage of the unique trend-tracking resource that the birding community represents. Checklist data can be used as an indicator of population trends in remote areas that lack other monitoring programs, and can be used in combination with other data where multiple monitoring programs coexist.
The following recommended guidelines for checklist programs were developed with these background considerations in mind.
1. The primary aim should be to collect data from relatively intensive general birding.
Casual records are also valuable and should be accepted (e.g. records of rare species seen on non-birding trips), but should be separable in the data base from more thorough birding records. Data collected for specialized surveys (e.g. hawk counts, tallies of birds at feeders), should be submitted to organizers of those projects first, but can also be contributed to a checklist program.:
Justification(a) Analyses of species richness and population trends require that "zeros" for a species indicate that the species was, in fact, not present. Therefore records limited in species coverage must be identifiable in the data base so that they can be eliminated from certain analyses (see detailed recommendations below). (b) Time needed for data base management and sorting for analysis can become prohibitive, so limits may have to be placed on what is accepted.
2. Each list should cover only one locality.
We recommend a flexible definition that allows each birder a bit of discretion, for example: "A single birding locality that can be traversed on foot in about an hour or less, usually separated from other such sites by some travel; i.e. an area not more than about one minute of latitude and longitude (3.2 km2, or 1.2 mi2); for example, a woodlot, a reservoir, or a park."
Justification for single locality:(a) A limited number of birds can be found in one locality, no matter how much effort is exerted, so judicious choice of a definition for "locality" should serve in a minor way to standardize effort. (b) This limitation should reduce any bias caused by birders covering more and more ground to find increasingly rare species. (c) Records can be used over the long term to track changes in bird fauna at specific birding destinations. (d) Bird records tied to locality can be used to identify sites that have unusual species or that are especially species-rich.
Justification for flexible definition:Alternatives are to define "locality" as an area of a certain size, a latitude- longitude block, or the area within a township (or other political unit). Political units are so unequal in size as to provide no uniformity among regions. It is frequently difficult to judge size of an area or to know where political or latitudinal boundaries fall. A birder whose area crosses boundaries or is slightly larger than the specified size will not want to split records into two separate lists. Therefore such definitions are likely to be ignored, and may discourage participation.
3. Each list should cover only one day's observations.
Justification: (a) Timing of migration is a typical use of checklist data, and resolution to day (rather than to week) allows detection of annual shifts in timing, and of early or late records. (The same applies to timing of breeding; see "Comments"). (b) Dates can be important for certain types of data selection (e.g. selecting equal samples from each date in order to avoid biases toward weekends, or selecting data restricted to a species' migration period). (c) Data recorded in units of days can readily be pooled by week or season if desired, but the reverse is of course not possible.
4. Observations should be recorded on a standard printed form (or in a standard electronic format).
Justification:(a) Ease of data entry and addition to data base. Anyone who has entered historic data into a data base knows the advantages of a standard format. (b) Existence of an official project list encourages participation, adherence to project guidelines, and submission of data.
1. Day_______ Month________ Year_________.
See justification for single day records, above.
2. Locality: Name nearest town or prominent map feature whose latitude and longitude can be looked up: ________________________ . County/Township___________. State/Province________.
Locality name (e.g. Smith Marsh, Jones Woods):____________.
Latitude/longitude to nearest minute (optional except for remote areas, but encouraged for regularly-visited sites): Latitude_______________. Longitude____________________.
See justification for single locality, above. Latitude and longitude are the locality identifiers that should be in the data base, and can be obtained from gazetteers, certain geographic software or topographic maps. Our suggested wording indicates what level of specificity is required in naming the locality. (Organizers of volunteer projects quickly learn that people have a genius for misinterpretation, so wording must be as specific as possible). State/ province need not be included if already printed on the form (e.g. in project name or address) and if the list is used only in one jurisdiction. County or township should always be included, however, because there may be many lakes etc. with the same name in different parts of a state.
3. Observer name(s)_________________________________________. Observer code (if known)_________________________________.
Each observer is assigned a number to be entered into the data base, and regular contributors can be asked to use these codes so that compilers do not have to keep looking them up. (In Quebec, 10% of the observers contribute 90% of the data.)
Justification:(a) Observer names are often needed for citation in status reports or sightings summaries, and for rechecking unusual records. Observer names have historical value. (b) We may want to know the number of observers contributing to a day's sample, or the number of checklists submitted per observer. (c) If names are entered instead of codes, there can be confusion (e.g. is "J. Smith" the same person as "John Smith"? Are all John Smiths the same person?).
4. Number or best estimate of each species detected. For estimates (e.g. of large flocks), record mid-point of probable range (in parentheses if especially uncertain). Underline the name of the species as you record the number seen.
[Each species found in the program's coverage area should then be listed in taxonomic order, and a space provided for recording the number detected. A few blank lines should be provided for write-ins. Species requiring extra documentation can be marked with an asterisk or other symbol.]
Justification for recording numbers:(a) The alternatives are to tick presence only, or to record numbers in categories (e.g. 1, 2-10, 11-100, 101-1000, 1000+). However, analysis has shown that recording of presence alone can mask a great deal of population change that is readily detectable if numbers are reported, even if of limited accuracy (Dunn and Larivée in prep., Bart and Klosiewski 1989). If numbers are thought to be too inaccurate, they can later be converted to presence/absence for analysis; whereas the reverse in not true. (b) Numbers can be adjusted for effort (e.g. converted to birds/hr), while presence and most categorical records cannot. (c) Coordinated checklist programs serve an educational role. Project reports, etc. should teach that an estimate of bird numbers is far more useful than a simple check mark, and that users are interested in order-of-magnitude changes that do not depend on total accuracy in counts.
Justification for specifying how to record estimates:(a) If observers feel they have to be exact, they may not participate, and allowing parentheses to indicate "guesses" for large numbers detected should increase comfort level in taking a stab at an estimate. (b) Specifying how to record estimates makes results more interpretable, and prevents people making up their own rules. For example, "500+" might mean "at least 500 but maybe 1000" (750 using our scheme), or it might mean "500-525" (512 with our scheme).
Justification for underlining species name:Lowers the chances of writing the number seen in the column for a different species.
5. Other information.
(a) Start time (to nearest quarter hour) _______. End time_______.
(b) Check one: This list reports birding that was general ______, or limited to one or a few species at the site ______ (e.g. drive-by sighting, waterfowl or feeder count).
(c) Check one: Ability of observer (or group) to detect and identify all species present (taking hearing into account) was: fair ______ good _______ excellent ______.
(d) Check one: Weather conditions for detecting birds in the habitat(s) visited was: fair ______ good _______ excellent ______.
Justification:(a) Start and end times allow correction of bird numbers to birds per hour, which helps standardize for variation in effort among counts. It also permits analysts to select data for counts of a given duration or that cover certain times of day. (b, c and d) Many analyses can use all available records, but certain others will be improved by excluding some lists (e.g. those with limited coverage, beginning observers or poor conditions for observation). Asking participants to "check one" helps reduce the number of multiple checks, which complicates data entry. (b) Careful wording needed. Some people will wonder if they should check "limited coverage" because the habitat visited only contains a limited set of species (e.g. seashore).
[Note: Collecting details of weather is not recommended because it is not necessary for most analyses and is tedious to record and manage in the data base.]
6. Comments: Note unusual or noteworthy behavior or plumage, evidence of breeding, etc. Provide documentation for rarities (append extra pages as needed).
Justification:Comments can provide much valuable, additional information that can readily be incorporated into the data base (see below). Extreme rarities should not be accepted without documentation.
1. Evidence of breeding.
Justification: For regions that have very little data on breeding status, a column could be added to the species list for breeding evidence codes. (Definitions should be printed on each list). For most regions, though, a column for breeding evidence will require effort by observers and compilers that may never be used, and the column will often be ignored. (In most cases, an absence of breeding codes will indicate a failure to keep records, rather than a lack of breeding evidence.)
2. Habitat(s) visited (check as many as apply): Deciduous woodland____ Coniferous woodland____ Scrub____ Grassland____ Agricultural____ Rural____ Suburban____ Urban____ Freshwater____ Salt water____ Other__________________________.
Justification: The list should be tailored to each project's coverage area. Listing the most common habitats reduces the number of new codes that will have to be devised for coding "other" habitats in the data base, and gives the observer the idea of what level of detail is desired. It might be possible to ask for estimated percent of each habitat at the site visited (e.g. 80% agricultural, 20% dediduous woodland).
The most likely use of habitat data would be in specialized analyses (e.g. trends in grassland birds might be studied using only lists that covered that habitat.) However, analysts interested in such questions are likely to use more standardized data sources. Most birding locations are not uniform, and when the observer visits more than one habitat, no analysis is possible of habitat- specific associations.
Unusual habitat associations are better documented by comments. Habitat data might be used to document change in habitat at popular birding destinations over time (but this is perhaps as easily learned from other sources). In short, project organizers should have clear ideas on how they wish to use habitat or other "optional" data before adding such data fields to their checklists.
(This is mostly a list of tasks that need to be taken care of by organizers, rather than recommendations per se.)
1. Provide volunteers with recording forms and instructions.
Project name, and address for further information and data submission, should be printed on every checklist. Emphasize the need to fill in all information on very card, as there are good reasons for each item being there. Brief instructions for volunteers should be printed directly on each checklist, but expanded instructions can be distributed separately (e.g. with each packet of lists).
2. Promote participation, and provide feedback on any results.
The value of a cooperative checklist program increases with its size, both in terms of geographic spread and in numbers of contributors. Feedback fosters continuation of participation, attracts new people, and fulfils an educational role. Feedback also requires regular examination of the data and evaluation of its quality, ensuring that problems are caught early and corrective action taken (e.g. clarifying instructions, or deleting incorrect records). Promotion and feedback can be done via news media, magazine articles or newsletters of the sponsoring organization and other natural history groups in the region.
3. Enter data into computer, in standardized format.
Original data entry might be done by the individual observer, by regional compilers, or centrally. Whoever does the job must be provided with all instructions on how to code and format the data. The central data base manager might do additional coding after receiving files from elsewhere (e.g. translating locality names to latitude and longitude, or replacing observer names with codes).
The authors are not in a position to recommend a particular data-entry or data management system, but there is plenty of software available that should prove suitable. We do strongly recommend, however, that project organizers work together to take advantage of one another's experience and perhaps to settle on a common data-storage format. The less translation of codes, sorting, reformatting etc. that must be done to pool data sets, the more likely that the full potential of checklist data will be tapped.
A few comments on Quebec's ÉPOQ system may provide some useful pointers.
1) Local bird clubs sign a contract agreeing to send data to the central compiler twice a year, in a standard format.
2) Clubs enter the data and submit it to the central compiler on diskettes. The average time for inputting one checklist is 3 minutes; one for preparation, two for typing. Many clubs have volunteers who specialize either in coding or in keyboarding.
3) Each checklist is given a unique number. This is written on the originals (which are stored in sequence for archival reference), and is entered into the data base to allow cross-reference among files.
4) Observer codes are in the form: first 3 letters of last name, first letter of first name. [A number could be added to such codes so that duplicates could be readily distinguished; e.g. DUNE001 and DUNE002].
5) The number code for each species is printed right on the checklist, for easy data entry. (This could be AOU number, or a number unique to each checklist project.)
6) Comments are entered into the database as codes. The first number refers to a class of comments (such as "breeding data", "injuries", "documentation of identification", "behavior", etc.). Then come numbers that refer to specific comments under that category (e.g. "nest with eggs", "recently-fledged young", etc.). ÉPOQ has created nearly 200 codes overall.
4. Edit lists, at least to detect obvious errors.
This can be done by local compilers or centrally, either before entry into the computer or afterwards. A computer program could be written to flag records with impossible dates or locations, or to flag records for birds that are extremely rare, unexpectedly abundant, or out of season. These can then be checked for transcription errors.
5. Develop a plan for archiving data.
This means assuring safe storage and filing for original checklists, so that a particular one can be found easily. It also means making periodic backups of the electronic data files, and transferring them when needed to more modern storage devices.
Many people read over drafts, and the final version was improved by comments from Pete Blancher, Greg Butcher, André Cyr, Brenda Dale, Connie Downes, Sam Droege, Jacques Larivée, C.J. Ralph, and Robert Rolley.
The authors of these guidelines would like to hear from anyone who has comments or wants more information (see email address on cover page).
1. We welcome your comments or criticisms on the recommendations themselves.
2. If you are involved in organizing a checklist project (existing or planned, for any organism), we'd be glad to hear about it, and we can put you in touch with other project organizers.
3. We would appreciate any advice from people who have experience in managing a checklist database, with your views on how these are most efficiently and effectively handled. We would also like to collect up information on data entry and management software that has been found useful for checklist projects.
4. Let us know if you want further information on the Migration Monitoring Council, or on recommended procedures for intensive migration counts (e.g. at bird observatories or banding stations).
5. We can provide you with a list of the codes used to summarize comments on checklists in Quebéc.
Direct comments or queries to Erica Dunn (firstname.lastname@example.org), Canadian Wildlife Service, National Wildlife Research Centre, 100 Gamelin Blvd., Hull, Quebec, Canada K1A 0H3.
Additional copies of this document are available from the same source, or from Greg Butcher, (email@example.com) American Birding Association, P.O. Box 6599, Colorado Springs, CO 80934-6599.BACK